pairwise does not work for corkendall or corspearman

[PGS62]

Currently pairwise works with cor but not with corkendall or corspearman

I think the fix is straightforward: Change the type declarations for each method of corkendall and corspearman to be AbstractVector in place of AbstractVector{<:Real} and AbstractMatrix in place of AbstractMatrix{<:Real}.

If I made a PR for that would it be likely to be merged?

julia> x = [1 2;3 4;5 missing]
3×2 Matrix{Union{Missing, Int64}}:
 1  2
 3  4
 5   missing

julia> pairwise(cor,eachcol(x),skipmissing=:pairwise)
2×2 Matrix{Float64}:
 1.0  1.0
 1.0  1.0

julia> pairwise(corkendall,eachcol(x),skipmissing=:pairwise)
ERROR: MethodError: no method matching corkendall(::SubArray{Union{Missing, Int64}, 1, Matrix{Union{Missing, Int64}}, Tuple{Vector{Int64}, Int64}, false}, ::SubArray{Union{Missing, Int64}, 1, Matrix{Union{Missing, Int64}}, Tuple{Vector{Int64}, Int64}, false})
Stacktrace:
 [1] _pairwise!(#unused#::Val{:pairwise}, f::typeof(corkendall), dest::Matrix{Float64}, x::Vector{SubArray{Union{Missing, Int64}, 1, Matrix{Union{Missing, Int64}}, Tuple{Base.Slice{Base.OneTo{Int64}}, Int64}, true}}, y::Vector{SubArray{Union{Missing, Int64}, 1, Matrix{Union{Missing, Int64}}, Tuple{Base.Slice{Base.OneTo{Int64}}, Int64}, true}}, symmetric::Bool)
   @ StatsBase C:\Users\phili\.julia\packages\StatsBase\XgjIN\src\pairwise.jl:63
 [2] _pairwise!(f::Function, dest::Matrix{Float64}, x::Vector{SubArray{Union{Missing, Int64}, 1, Matrix{Union{Missing, Int64}}, Tuple{Base.Slice{Base.OneTo{Int64}}, Int64}, true}}, y::Vector{SubArray{Union{Missing, Int64}, 1, Matrix{Union{Missing, Int64}}, Tuple{Base.Slice{Base.OneTo{Int64}}, Int64}, true}}; symmetric::Bool, skipmissing::Symbol)
   @ StatsBase C:\Users\phili\.julia\packages\StatsBase\XgjIN\src\pairwise.jl:121
 [3] _pairwise(#unused#::Val{:pairwise}, f::typeof(corkendall), x::Base.Generator{Base.OneTo{Int64}, Base.var"#242#243"{Matrix{Union{Missing, Int64}}}}, y::Base.Generator{Base.OneTo{Int64}, Base.var"#242#243"{Matrix{Union{Missing, Int64}}}}, symmetric::Bool)
   @ StatsBase C:\Users\phili\.julia\packages\StatsBase\XgjIN\src\pairwise.jl:172
 [4] pairwise(f::Function, x::Base.Generator{Base.OneTo{Int64}, Base.var"#242#243"{Matrix{Union{Missing, Int64}}}}, y::Base.Generator{Base.OneTo{Int64}, Base.var"#242#243"{Matrix{Union{Missing, Int64}}}}; symmetric::Bool, skipmissing::Symbol)
   @ StatsBase C:\Users\phili\.julia\packages\StatsBase\XgjIN\src\pairwise.jl:315
 [5] top-level scope
   @ REPL[17]:1

[PGS62]

Closing because I think the solution proposed is not a good one. Better to change pairwise so that it calls f passing in x and y of type AbstractVector{<:Real}.